Cohort analysis in mobile applications: how to understand that the economy is converging?

Cohort analysis Is a way to track the “density” of a metric for specific user groups. Unlike regular graphs, cohorts show exactly the stickiness of users, distributed over time.

It sounds difficult, but in reality it is easier to understand. Let’s take a concrete example: let’s try to understand how the subscriber base behaves in a mobile application: how they pay, unsubscribe and how long they live. I will show the answers to these questions using the example of our implementation in Adapty

An example of cohort analysis implementation from Adapty
An example of cohort analysis implementation from Adapty

When constructing cohorts, it is important to indicate how they are formed, what to display in cells, and what to measure. In our case, cohorts are formed by the month of installation, in the cell we show the revenue, the number of subscribers and the average revenue per subscriber for each month (ARPPU).

That is, in the lines we have a cohort (group) of users who installed the application in a certain month, and in the columns – the values ​​of the metrics for each month from the month of installation (this is the first month or M1). In the highlighted line, M1 is January, M2 is February, etc. Every month the number of subscribers in this cohort does not increase: even if the user installed in January and subscribed in February, he will be assigned to the cohort in January. It seems to us that this method is correct for evaluating convergence:

The idea of ​​evaluating the economy this way is about how user acquisition works. When purchasing advertising, the developer pays for installations one way or another, and not for targeted actions. Even in a CPA campaign, everything will be related to the cost of installation (CPI). Therefore, in order to assess the effectiveness of buying traffic, you need to look at how exactly people who installed the application during this period will be monetized. At the same time, if the user has installed the application, but has not paid for a month, he will only get into M2.

We see that the January cohort of users has brought us a total of $ 2900 in revenue from 73 subscribers up to the current moment:

Total revenue from 73 subscribers in 4 months - $ 2.9K
Total revenue from 73 subscribers in 4 months – $ 2.9K

Next, we see the dynamics of the cohort decay: how quickly users unsubscribe. In January we had 67 subscribers, in February we already had 38, and as of June there were 10 left (a shaded cell means that subscribers can still be added, since the month is not over).

Will the economy converge?

Now, suppose we paid $ 4000 for advertising in January, our question is “will users pay off or not?”, That is, will the revenue from them be more than $ 4000 in a reasonable time.

Let’s take a closer look at the dynamics of subscribers in the January cohort.

January cohort by month (first column - settings)
January cohort by month (first column – settings)

January cohort by month (first column – settings)

At the moment, the application has earned $ 2900 before Apple commission, or $ 2,465 after deducting 15% (the app is in Apple’s SMB program). We will also assume that we sell weekly subscriptions for an average of $ 10.

We see that the number of active subscribers after the first month fell by almost 2 times, then by 20%, then by only 10% to 24. Since at the time of writing the notes, the month has not yet ended, let’s take the best scenario – let all 24 subscribers be with us always. And even if they pay $ 15.48 further on average, that is, their ARPPU does not change.

To get to $ 4,000, you need subscribers to pay more than $ 1,500. Even with a monthly revenue of $ 372 and zero unsubscription, the cohort will converge at best after 4-5 months of continuous payments. In practice, taking into account the previous dynamics, and knowing that traffic is bought evenly, the cohort is unlikely to converge in less than a couple of years, but in fact, it will most likely be at a loss. The reason is that weekly subscriptions constantly remind of themselves, and users are less likely to remain in long-term payments, because if the application is good, it is much more profitable to buy a year. But even with monthly subscriptions with such dynamics, one can hardly expect a positive profit.

The convergence of the subscription economy is a long-term process. We at Adapty have made a tool to test if the economy will converge on your numbers – Subscription Calculator… Try it to evaluate how to make money on mobile applications.

Similar Posts

Leave a Reply